The EMBO Journal Vol.19 No.5 pp.921–930, 2000

Camel heavy-chain antibodies: diverse germline VHHand specific mechanisms enlarge the antigen-binding repertoire

(Muyldermans et al., 1994; Vu et al., 1997). Most strikingViet Khong Nguyen1, Raymond Hamers,are the amino acid substitutions V37F (Val at position 37Lode Wyns and Serge Muyldermansin the VH to Phe in the VHH), or V37Y, G44E, L45R or

Department Ultrastructure, Vlaams Interuniversitair Instituut voor L45C, and W47 most often to G [numbers refer to theBiotechnologie, Vrije Universiteit Brussel, Paardenstraat 65, amino acid positions numbered according to Kabat et al.B-1640 Sint Genesius Rode, Belgium

(1991)]. In the conventional VHs, these FR2 amino acids1Corresponding author interact with the variable domain of the light chain (VL),e-mail: nguykhon@vub.ac.be and are conserved during evolution (Kabat et al., 1991).

The CDR3 of the VHH is longer on average than that ofThe antigen-binding site of the camel heavy-chaina VH domain (Vu et al., 1997), and is often constrainedantibodies devoid of light chain consists of a singleby an interloop disulfide bond (Davies and Riechmann,variable domain (VHH) that obviously lacks the VH–1996; Desmyter et al., 1996).VL combinatorial diversity. To evaluate the extent

A high titre and a complex repertoire of HCAbs canof the VHH antigen-binding repertoire, a germlinebe obtained from immunized or infected dromedaries ordatabase was constructed from PCR-amplified VHH/llamas (Hamers-Casterman et al., 1993; Ghahroudi et al.,VH segments of a single specimen of Camelus1997). Many HCAbs raised against enzymes are competit-dromedarius. A total of 33 VHH and 39 VH uniqueive inhibitors (Lauwereys et al., 1998). This is surprising,sequences were identified, encoded by 42 and 50since the active site of enzymes has a low antigenicity fordifferent genes, respectively. Sequence comparisonconventional Abs (Novotny, 1991). Thus, the HCAbsindicates that the VHHs evolved within the VH subgrouprecognize a broad range of epitopes, some of which differIII. Nevertheless, the VHH germline segments arefrom those for conventional Abs.highly diverse, leading to a broad structural repertoire

Previously, we identified germline VH and VHH segmentsof the antigen-binding loops. Seven VHH subfamiliesindicating that the variable domain of the HCAbs iswere recognized, of which five were confirmed to beencoded by a distinct set of V genes (Nguyen et al.,expressed in vivo. Comparison of germline and cDNA1998). In this study, we investigate the potential VHHsequences demonstrates that the rearranged VHHs aregermline repertoire to gain insight into the ways by whichextensively diversified by somatic mutation processes,the dromedary HCAbs acquire a complex repertoire ofleading to an additional hypervariable region and aAg-binding sites. In conventional Abs, the diversity ofhigh incidence of nucleotide insertions or deletions.the Ag-binding site is generated at multiple levels. TheThese diversification processes are driven by hyper-VH is generated by assembling variable (V), diversity (D)mutation and recombination hotspots embedded inand joining (J) elements (Tonegawa, 1983), in which thethe VHH germline genes at the regions affecting theV-gene segment encodes the CDR1 and CDR2; the CDR3structure of the antigen-binding loops.is generated by the V–D–J joining. In this joining process,Keywords: antigen-binding repertoire/camel heavy-chaingreat sequence variation is introduced by non-templateantibody/gene replacement/germline VH /hypermutationaddition of nucleotides at the V–D and D–J junctionshotspots(junctional diversity). Random association of a VH and aVL (combinatorial diversity) generates an immenselydiverse Ag-binding repertoire. Additional diversification

Introduction of the Ag-binding repertoire could be achieved by somatichypermutation (Berek et al., 1991) and gene conversionThe emergence of the Camelidae (camels and llamas)(Reynaud et al., 1987; Becker and Knight, 1990). Thus,within the Artiodactyls is accompanied by a unique eventthe primary Ag-binding repertoire of the HCAbs lackingin immunoglobulin (Ig) evolution, namely the appearancethe VH–VL combinatorial diversity relies on the innateof additional classes of functional antibodies (Abs) com-number and sequence diversity of the VHH germlineposed solely of heavy chains (Hamers-Casterman et al.,segments and the junctional diversity.1993). These heavy-chain antibodies (HCAbs) lack the

The identification of the germline VHH genes is notfirst domain of the constant region (CH1), which is presentonly of fundamental interest but also has a potentialin the genome but is spliced out during mRNA processingbiotechnological benefit. At the moment, HCAbs with(Nguyen et al., 1999; Woolven et al., 1999). The antigenenzyme inhibiting activity can only be obtained after(Ag)-binding site of these HCAbs is composed of a singleimmunizing camels or llamas. Techniques have beenvariable domain (referred to as VHH). The VHH structuredeveloped to retrieve various binders from syntheticresembles that of the heavy chain variable domain (VH)libraries of Ab fragments (Hoogenboom and Winter, 1992;of the conventional Abs. However, there are remarkableWinter et al., 1994). Single-domain Ab libraries have beensequence differences at the second framework (FR2) and

the third complementarity-determining region (CDR3) constructed by adding a synthetic CDR3 region to the

© European Molecular Biology Organization 921

V.K.Nguyen et al.

from liver DNA of a single dromedary. Out of255 sequences we found 145 different sequences, whichformed an initial germline VH/VHH database. Assuming aPCR error rate of 2�10–4 errors/base (Cha and Thilly,1993), we estimated that up to four base errors mighthave accumulated within each 600 bp sequence. Hence,we only considered clones differing by at least 5 nucleo-tides (nt), which reduced the VH/VHH germline databasefrom 145 to 94 sequences. The nucleotide sequences ofthese 94 clones have been submitted to the DDBJ/EMBL/GenBank databank (accession Nos AJ245107–AJ245200).

In an alternative approach, we screened the dromedarygenomic library cloned in phage λ (Nguyen et al., 1998).A total of 55 different phage clones were isolated andtheir VH or VHH regions were subcloned and sequenced.However, no new VH/VHH sequences were discovered,supporting that our database of 94 VH/VHH sequencesobtained by PCR is representative to evaluate the potentialVHH repertoire.

Fig. 1. Southern blot analysis of dromedary liver genomic DNAhybridized with a camel VHH probe. Dromedary DNA was digestedwith BamHI, EcoRI and BamHI–EcoRI, as indicated on top of the

Characterization of VH/VHH sequences and VHHlanes. Phage λ PstI restriction fragments are used as size marker.subfamilies

All 94 genomic VH/VHH sequences span from the con-known human VH elements (Davies and Riechmann, 1995;served octamer to the Cys92 of FR3. They contain aReiter et al., 1999). It would be an asset if similar librariesTATA box 105 nt upstream of the ATG initiation codon.of VHHs were available to retrieve binders or inhibitorsThis ATG starts an open reading frame (ORF) of conservedof interest. To be as successful as the dromedary in makingsequences encoding a leader signal peptide, which isgood Ag binders we should start from a library that coversinterrupted by an intron of 104 bp. The remaining sequenceall potential VHH segments. In addition, analysis of theencodes the VH or VHH segment. Two sequences (cvhp51amino acids that are mutated during the in vivo affinityand cvhp52) are interrupted by stop codons. These clonesmaturation would provide a rational strategy for increasingcould be pseudo-genes and were excluded from furtherthe repertoire of the VHH library or to improve the affinity

of binders. analyses. The remaining 92 clones probably encode func-We cloned from a single dromedary the germline tional genes as no aberrant amino acid occurs in the ORF

VHH gene segments to analyse their complexity. The at the critical positions for the Ig fold (Chothia et al., 1988).comparisons of the germline and cDNA VH/VHH sequences

Identification of VH/VHH sequences. The nucleotide iden-reveal the somatic diversification mechanisms used by thetity between any pair of the 92 VH/VHH encoding sequencescamelids to enlarge the primary Ag-binding repertoire ofis �80%. The amino acid sequences of FR1 and thethe HCAbs. The involvement of DNA signal sequencesbeginning of FR3 are homologous to each other (Figure 2)in these diversification processes is discussed.and to the human or mouse VH of family 3. Therefore,according to the family definitions (Brodeur and Riblet,Results1984; Schroeder et al., 1990), all these camel sequences

Southern blot analysis of the genomic DNA could be assigned to the VH family 3. Based on crucialA rough estimate of the VH/VHH germline repertoire was amino acid differences in the FR2 region, we havefirst obtained by Southern blot analysis of dromedary liver previously distinguished the camel VH3 family (used inDNA, probed by the PCR fragments from the upstream the conventional Abs) and the VH3H family (for theconserved octamer sequence to the FR3 of camel germline HCAbs) (Nguyen et al., 1998). The current database canVH or VHH clones (Nguyen et al., 1998). The hybridization also be divided into these two families according to thepatterns shown by these two different probes were ident- presence of V37, G44, L45 and W47 (VH3), and F/Y37,ical. In single digestions with BamHI or EcoRI, 11 bands Q/E44 and R/C45 (VH3H) (Figure 2). Out of 92 VHs/of different intensity and ranging from 1.7 to �14 kb VHHs, 50 sequences were clearly assigned to VH3 (cvhp01–could be visualized (Figure 1). The double digest revealed cvhp50) and 42 to the VH3H family (cvhhp01–cvhhp42).at least 15 bands. Since we never observed an EcoRI or This VH/VHH ratio (1.2/1) was also found among sequencesBamHI restriction site between the octamer and FR3 obtained from the phage λ-isolated clones. We thereforesequences of VH/VHH (this work), we conclude that the conclude that a similar number of VH3 and VH3H genesdromedary VH/VHH germline gene segments are spread

reside in the dromedary genome.over a minimum of 15 different EcoRI–BamHI sizeComparison of amino acid sequences revealed that afamilies.

number of clones have an identical coding capacity (boxedin Figure 2) although their nucleotide sequence differedDetermination of dromedary germline VH/VHHby �4 nt. As a result, the 50 VH and 42 VHH genesequencessegments code respectively for 39 VH and 33 VHH uniqueA germline V-gene database was obtained by cloning and

sequencing PCR fragments of VH/VHH elements obtained amino acid sequences.

922

Camel VH and VHH germline genes

VHH subfamilies. Multiple alignment of the VHH amino CDR1) and R71Q dictating the conformation of the H2loop (the exposed part of the CDR2).acid sequences revealed the clustering of sequences with

an additional cysteine at position 30, 32, 33 or 45, and From crystallographic data on the VHH structures, weinfer that the combination Y27–Y29 (found in 22 germlinewith a CDR2 length of 16 or 17 amino acids (Figure 2).

Based on these hallmarks, we further classified the 42 VHHs) can adopt either a type-1 or type-4 canonicalstructure, depending on the nature of the residue 31clones of the VHH into seven subfamilies named 1a, 2a,

2b, 3b, 4b, 5a and 5b (Table I; Figure 2). (Decanniere et al., 1999). Similarly, the six-residue H2loops associated with R71 reveal either a H2 type-2A neighbour-joining tree was constructed using the

VH/VHH coding sequences to analyse the VH and VHH (Spinelli et al., 1996) or a novel H2 structure (Decanniereet al., 1999). These loops are therefore denoted as ‘PS’relationship. The dendrogram (Figure 3) shows clearly

two clusters, one cluster contains only the VH3 and the in Table II for ‘potentiality to switch’.Thus, the dromedary VH germline sequences have aother the VH3H genes, despite the fact that the pairwise

sequence identity among VHs or among VHHs may be structural repertoire of six different H1–H2 combinations(Table II), whereas the structural repertoire of the germlinelower than that of a VH–VHH pair. Moreover, the branches

appear to correspond to the VHH subfamilies defined above. VHHs is far more diverse, up to 10 different combinationsof H1 and H2 loop conformations are observed. Moreover,it seems that both the H1 and H2 loop conformations ofDetermination of expressed VHH gene segmentsthe VHHs are apt to reshape due to minor modifications,We assembled a VHH cDNA database containing 103thereby further enlarging the VHH structural repertoire.sequences, 32 of which have a known Ag specificity.

Thirty-one of the 103 cDNAs have a length deviatingVariability of camel VH/VHH sequencesfrom any of the germline VHH segments. While theseThe variability plots (Kabat et al., 1991) were constructedclones could originate from a germline VHH that has notfor the germline VHs and VHHs (Figure 5, upper panels).yet been discovered, they could also originate from unequalBoth histograms delineate the conventional CDR1 andDNA recombination or gene conversion (Reynaud et al.,CDR2 regions as the sites of greater variability.1987). The remaining 72 VHH cDNA sequences (26 with

known Ag specificity) were compared with all 42 germline Hypervariability in the H1 region of VHHs cDNA. TheVHH sequences to reveal the closest sequence identity. variability of the VH and VHH cDNAs was plotted usingThe combination of the sequence identity score and the sets of 50 VH and 42 VHH cDNAs, and compared withtwo VHH subfamily hallmarks defined earlier leaves no that of the germline sequences (Figure 5). The overalldoubt as to the identification of the germline subfamily. variability is much higher in cDNAs than in the germline,Among the 72 VHH cDNA clones, there is no indication indicating that somatic mutation plays an important rolefor the in vivo expression of members of subfamilies 1a in the generation of the camel VH/VHH repertoire. Theand 5a (Table I). The result shows that at least five most striking observation is the presence, exclusively inout of the seven VHH subfamilies are used to generate the VHH cDNAs, of an additional hypervariable regionproductive VHH domains found in HCAbs. (residues 27–30) located upstream of the conventional

CDR1 region.Identical D and JH segments in VH and VHH cDNAs. Agermline D mini-gene (similar to human D4) was identified Hypermutational hotspots imprinted in the germline.with an RSS at both ends. We noted that some VH and Somatic diversification results from different mechanismsVHH cDNAs employ this D element (Figure 4). In addition, including gene conversion, and somatic hypermutationour cDNA database also contains VHs and VHHs that make (Wagner and Neuberger, 1996; Neuberger et al., 1998).use of the same JH element. This would indicate that VH The latter was proposed to be driven by hypermutationaland VHH genes are within the same functional V region hotspots such as the AGY and TAY (Y � C or T)of the dromedary genome. sequences (Yelamos et al., 1995; Milstein et al., 1998).

The occurrence of these hotspots in the germline VHs andVHHs was superimposed on the cDNA variability plotsStructural repertoire of Ag-binding loops of camel(Figure 6). Clearly, the occurrence of hotspots correspondsgermline VH/VHHto the highest variability at the CDR1 and CDR2 of cDNAThe structure of the Ag-binding loops of the camelsequences. The hypervariable region found exclusively ingermline VHs and VHHs was predicted by algorithmsthe VHH cDNA (residues 27–30) could result from theof Chothia and Martin (http://www.biochem.ucl.ac.uk/TAY hotspots that are present in the germline VHHs, but~martin/abs/chothia.html) (Chothia et al., 1989, 1992;absent in the germline VHs.Martin and Thornton, 1996), and by taking into account

the crystallographic structures of dromedary and llamaVHHs (Desmyter et al., 1996; Spinelli et al., 1996; Putative recombination

As mentioned previously, the 31 VHH cDNAs havingDecanniere et al., 1999) (Table II; Figure 2).The Ag-binding loops of the dromedary VHs are pre- lengths deviating from that of the germline could originate

from unequal DNA recombination, gene conversion ordicted to conform to the known canonical structurescharacterized by well-defined key amino acids. In contrast, gene replacement. Comparison of the nucleotide sequences

of these off-size cDNA with germline VHHs revealed fourthe Ag-binding loops of the VHHs are expected to deviatefrequently from the known canonical structures, mainly distinct regions in which nucleotides were inserted or

deleted (Ins/Del). The four regions surround the residuesdue to the substitution of key amino acids. Notablesubstitutions are G26E and F29S in the H1 loop (the 24 � 1 (1 Del and 5 Ins), 30 � 3 (9 Dels and 6 Ins),

54 � 3 (2 Dels and 8 Ins) and 74 � 1 (3 Dels). We notedsolvent-exposed part of the loop overlapping with the

923

V.K.Nguyen et al.

924

Camel VH and VHH germline genes

that most of these Ins/Del regions are within or at the and it also abuts a deletion at position 74–75 found inthree individual VHH cDNA clones, indicating a possibleborder of peculiar DNA sequences in the germline VHHs,

suggesting an active involvement of these sequences in causal relationship with the heptamer-like sequence inthese deletion events (Figure 7D).the Ins/Del events. For example, a palindromic sequence

‘CAGTAGCTACTG’ (corresponding to residues 30–33)borders the Del/Ins region 30 � 3 (example in Figure 7A).

DiscussionSimilarly, the Ins/Dels at region 54 � 3 appear to coincidewith the presence of another palindrome ‘CTATTAATAG’ Dromedary VH/VHH gene segments(codons 51–52a) (Figure 7B). Furthermore, two different To evaluate the potential Ag-binding repertoire of thetypes of nucleotide insertions can be discerned. Among dromedary HCAbs, we established representative data-17 insertion events, eight are clearly duplications of bases of the germline and cDNA VH/VHH sequences. Thethe bordering sequences, while nine are non-templated germline database contains 94 V-gene segments, of whichnucleotide insertions. An example of each case is shown the majority were found in two independent PCR experi-in Figure 7B. ments and confirmed by sequences derived from a genomic

The high incidence of the heptamer-like sequence, a library. Statistical analysis taking into account the totalcomponent of the Ig recombination signal (RSS) is another number of sequenced clones and the number of doubles,peculiar feature of the VHH sequence (Nguyen et al., triples, etc. also indicated that the dromedary may contain1998). The RSS occurrence is almost twice as high in the a total of 110 � 20 VH and VHH genes. This estimationgermline VHHs (37/42) as it is in the VHs (21/50) is compatible with the Southern blot analysis (Figure 1),(Figure 7C). The majority of these RSS in the VHH are in which the intensively labelled bands may correspondconcentrated at the FR3 site corresponding to residues to VH/VHH genes residing in multiple fragments of similar76–78. Interestingly, this RSS location is downstream of size that overlap in our blot, and the longer fragmentsa region with a modest increased variability (Figure 5), may contain more than one V element. The V-gene

segments in our database have a high degree of sequencehomology, several encoding an identical amino acidsequence. Some of these could be the result of a recentgene duplication, or alternatively, they might be allelicvariants as the sequences were derived from a diploidgenome. However, the allelic location of the genes haslittle relevance to our goal to evaluate the potential VHHrepertoire.

Clearly, the dromedary contains two distinct sets of Vgenes (~40 VHHs and ~50 VHs), encoding the V domainof the conventional Abs and HCAbs, respectively. The VHand VHH genes appear to be within the same functionalV region of the dromedary genome. The identification ofan identical D mini-gene in a VH cDNA and a VHH cDNA(Figure 4) suggests the common use of the D segmentsFig. 3. Neighbour-joining phylogenetic tree of the dromedary germlinefor VH and VHH. Moreover, preliminary evidence showsVH and VHH segments. The tree was constructed by using clustalW,

phylip packages with 1000 replicates for Bootstrap of nucleotide that rearranged VH and VHH genes bear identical 3�sequences encoding the VH/VHH portions. Filled and open circles sequences downstream of the DJH segments (unpublisheddenote a VH and VHH member, respectively. The two filled rectangles results). However, within this functional V region it isindicate pseudogenes. VHH subfamilies are indicated.

still unclear whether the germline VHs and VHHs areinterspersed or clustered.

Diverse germline VHHs

All germline VHHs reported here belong to a single family.However, the intrinsic structural repertoire of their Ag-binding loops is quite diverse. The prediction of thecanonical structure shows that 10 combinations of theH1–H2 loops are possible in the VHHs, which is higherFig. 4. The common usage of the D element in VH and VHH: a

genomic D element (camD4) is flanked at both sides with the RSS than what is observed in the VH of family 3, e.g. six incontaining a nonamer (N) separated from a heptamer (H) by a 12-nt the dromedary, five in human and three in mousespacer. The sequences derived from the germline camD4 (upper line), (Tomlinson et al., 1992; Almagro et al., 1997). A total ofVH-cDNA clone i5h50 (middle line) and VHH-cDNA clone med83

10 combinations of the H1–H2 loop structures were found(lower line) are boxed. The CDR3 region between the FR3 and FR4 isindicated. in human and mouse germline VHs of all families (Almagro

Fig. 2. The deduced amino acid sequences of the dromedary germline VH (upper) and VHH segments (lower). Braces group the VHH sequences insubfamilies as defined by their CDR2 length and the position of the additional Cys at indicated positions. Boxes indicate V genes that differ by atleast 5 nt, but encode an identical amino acid sequence. The amino acid length is given in the column following the primary structure. The types ofthe predicted H1 and H2 canonical structures are in the last columns. Asterisks denote where the predicted loop may deviate from loop type definedby Chothia et al. (1992) due to novel residue at the key-site. X, unpredictable; PS, potentiality to switch (see text).

925

V.K.Nguyen et al.

Table I. VHH germline sub-families and their cDNA counterparts

Subfamily Cys/CDR2a Members Total cDNA found Binder found Agsb

VH3H–1a Cys–/17 1 0 0VH3H–2a Cys33/17 5 54 18 lysozyme, carbonic and anhydrase, cutinase,

tetanus toxoid, RNase phenyl oxazoloneVH3H–2b Cys33/16 1 2 2 lysozyme, carbonic anhydraseVH3H–3b Cys30/16 2 5 2 amylaseVH3H–4b Cys45/16 5 8 3 amylase, tetanus toxoidVH3H–5a Cys32/17 2 0 0VH3H–5b Cys32/16 26 3 1 amylaseTotal 42 72 26

aThe presence of an additional Cys at indicated position/the length of CDR2 (in amino acids).bReference from Desmyter et al. (1996), Ghahroudi et al. (1997), Lauwereys et al. (1998), Decanniere et al. (1999) and other unpublishedAg-binders provided by M.Lauwereys, K.Silence, T.Laeremans and M.T.M.C.Serrao.

Table II. Structural repertoire of the camel germline VH and VHH segments

VH VHH

H1 H2 VHCSC members H1 H2 VHCSC members

24 26 34 52 71 24 26 34 52 71A GFTFSGY M Y-SDGG R 1–1 27 A GYTFSSY M I-SDGS R 1–1 5A GFTFSSY M Y-SDGS Q 1–1* 1 A GFT@@CS M S-TDGS K 1*-1 2A GFTFSSY @ NSDGSN R 1–3 14 A GYIFSCS M S-SDGS Q 1*-1* 3A GFTFSSY @ @TGGGS K/Q 1–2* 6 A GYTYSSC M @-@DG@ K PS-1 4A GFTSSSY M YTGGGS K X-2* 1 A GYTY@S@ M D-SDGS Q PS-1* 12A AFTYSSC M NSGGGS Q 1*-2* 1 A GFTSN@C M S-TDG@ K X-1 8

A GFTFSSY M NSGGGS Q 1–2* 1A GYIFSCS M NSGGGS R 1*-PS 1A GYTYSS@ M @@GGGS Q PS-2* 4A @YTYSS@ M @@GGGS R PS-PS 2

VHCSC or VHHCSC: canonical structure class of VH or VHH, respectively. Asterisks denote that the predicted loop may deviate from the loop-type asdefined by Chothia et al. (1992) due to a novel residue (underlined) at the key site. The @ denotes more than one residue occurring at that site.X, unpredictable; PS, potentiality to switch (see text).

et al., 1997). The higher complexity of the germline VHHs would generate a sufficiently diverse repertoire to retrievegood Ag binders.within one family is due to the presence of novel residues

at key sites for the loop conformation, the variable lengthof the CDR2 and the presence of an additional Cys located Somatic diversification of VHHs

From a limited number of germline V genes, many speciesat position 30, 32, 33 or 45.can generate a large Ag-binding repertoire through asomatic diversification process (Reynaud et al., 1989;VHH usage in the dromedary

A total of 42 VHH gene segments were identified, which Knight, 1992; Sun et al., 1994; Dufour et al., 1996;Sinclair et al., 1997). We note that the dromedary VHHencode 33 unique sequences. In contrast to rabbits where

one VH germline is predominantly rearranged (Knight, structural repertoire is readily diversified by the introduc-tion of an additional disulfide bridge, the high incidence1992), members of at least five out of seven VHH subfamil-

ies are used to generate HCAbs in the dromedary (Table of nucleotide Ins/Dels, gene replacement and the extensivesomatic hyperpoint mutations.I). It also appears that the larger subfamilies were not

necessarily the most frequently encountered in the cDNAExtra disulfide bridge. The germline VHH possesses a Cysclones. On the contrary, five members of subfamily 2a

accounted for 75% of the inspected cDNAs, while 26 in addition to the conserved Cys22 and Cys92 that forma disulfide bond in all Ig domains. This additional Cys issequences of subfamily 5b matched only three cDNA

clones (Table I). This preferential usage of particular VHH maintained in the vast majority of the VHH cDNAs,which also acquired an additional Cys within the CDR3segments could be due to their proximal location to the

D-gene cluster, as observed in other mammals e.g. the (Muyldermans et al., 1994; Vu et al., 1997). The disulfidebonds between these additional cysteines cross-links thehuman VH3–23 (Stewart et al., 1992) and mouse VH10

(Schiff et al., 1988). Ag-binding loops (Davies and Riechmann, 1996; Desmyteret al., 1996). This has two consequences: it stabilizes theWe note that the Ag-binding loop structures of the

VHH subfamily 2a are predicted to switch readily. This domain, and it induces constraints in the CDR1 or CDR3,which could lead to novel loop conformations and therebysubfamily appears to be the most prevalent to generate

Ag binders, indicating that the paratope repertoire derived increasing the paratope repertoire.from these VHH members could cope with various Ags.Therefore, the selection of these germline VHHs as a High incidence of insertions and deletions. The Ins/Dels

of nucleotides found in 31 VHH cDNAs cluster in fourscaffold to construct a synthetic single domain library

926

Camel VH and VHH germline genes

Fig. 5. Amino acid sequence variability plots for the VH and VHH germline (upper panels) and cDNA sequences (lower panels). The variability index(bars) was calculated from FR1 to FR3, as described in the results. The grey and open bars are for amino acid positions of the CDR and FR regions,respectively, and the black bars are for the extra hypervariable region in the VHH. The CDR1 and CDR2, as defined by Kabat et al. (1991), areshaded and placed between dotted lines.

Fig. 6. Distribution and frequency of AGY (filled bars) and TAY (open bars) triplets in the germline VH (upper) and VHH (lower) segments. Theoccurrence of AGY and TAY triplets were scored in all reading frames. The numbers in between refer to the position of VH codons (Kabatnumbering). The shaded background originated from the cDNA variability plots shown in Figure 5. The dotted regions denote the CDR1 and CDR2,as defined by Kabat et al. (1991).

regions. The high incidence (30%) and the clustering for chicken, rabbit and cattle (Reynaud et al., 1987; Beckerand Knight, 1990; Parng et al., 1996).feature of Ins/Dels suggest that they (at least in part)

could be relics of a gene conversion-like event as described The equal presence of two different types of nucleotide

927

V.K.Nguyen et al.

found in non-functional Ig genes (Klein et al., 1998), inthe dromedary, six out of 31 off-size cDNAs have aknown Ag specificity.

Gene replacement. While the exact Ins/Del mechanisminvolving palindromes remains unknown, the high incid-ence of the embedded RSS at VHH-FR3 abutting the Del/Ins region 74 � 1 [Figure 7D and Ins at position 75A oftwo llama VHHs (Vu et al., 1997)] supports the genereplacement mechanism (Reth et al., 1986) involvingrecombination-activating gene (RAG) proteins as observedin human and mouse (Kleinfield et al., 1986; Komoriet al., 1993). The involvement of a RAG-like activity toinitiate the gene replacement predicts a DNA cleavagetwo bases upstream of the heptamer sequence (Ramsdenet al., 1996). This is exactly the position where thedeletions are observed in these individual clones(Figure 7D). It is possible that improper joining (Meleket al., 1998) causes these Ins/Dels, and proper joiningretains the sequence length but induces nucleotide replace-ments (Weinstein et al., 1994). The latter possibly leadsto the increased variability observed in this region of theVHH (Figure 5). The region around position 75 lies in asolvent exposed loop immediately adjacent to the H2 loopin the folded Ig domain. In monomeric T-cell receptors,the corresponding region is also hypervariable and reportedto interact with ligand (Howell et al., 1991). By analogy,we suppose that this VHH region can be involved in the

Fig. 7. (A) Nucleotide alignment of cDNA clone l3a11 (lower line) Ag–Ab interaction by either directly contacting the Ag orand its putative germline gene (cvhhp11, upper line), revealing a inducing structural changes in the Ag-binding site. Takendeletion of 6 nt (dotted). The palindromic sequence is in bold and together, it is plausible that gene replacement at thisunderlined. Numbers indicate nucleotide positions of the germline

region, in which a new VHH rearranges to the pre-existingVHH element. (B) Alignment of the cDNA (lower lines) and theVHH-D-JH, results in a novel Ag-binding site composedclosest germline sequences (upper lines) indicates insertions, which are

non-templated (pair 1) or are duplicated (boxed in pair 2). of the incoming VHH and the existing CDR3.(C) Frequency and distribution of the RSS-like elements in the camel

Extended hypervariable CDR1 region. The patterns of thegermline VHs and VHHs. Numbers indicate the incidence of RSS foundin 50 VHs (upper) and 42 VHHs (lower). (D) Improper joining adjacent cDNA variability plots (Figure 5) are basically consistentto the RSS signal resulting in clones with a deletion: the heptameric with the classification of FR- and CDR-regions as definedsequences are boxed and the expected cleavage site is indicated by an

by Kabat et al. (1991). However, the extra hypervariablearrow. The three camel VHH cDNA sequences (l3a08, med64 andregion (residues 27–30) present exclusively in the VHH ishhat20) show a codon deletion when aligned with their putative

corresponding genes (cvhhp28 and cvhhp11). most remarkable. Crystallographic studies of VHH–antigencomplexes demonstrated that amino acids located in thisarea interact with Ag (Desmyter et al., 1996; Decanniereinsertions indicates that there are at least two differentet al., 1999). Obviously, the VHH uses this region togethermodes of action leading to these sequence lengthwith a long CDR3 to increase the surface area interactingvariations. Half of the insertions are nucleotide duplica-with Ags. Surprisingly, to attain new amino acids attions, indicating that these Ins/Dels could be by-productsthese positions the germline VHHs accumulated two newof somatic hypermutation (Wilson et al., 1998). In thehypermutation hotspots (Figure 6) by two single pointother half of the events, the inserted nucleotide sequencesmutations. The non-hotspot triplet TTY in the VH isare different from their flanking sequences. Therefore, thissubstituted by TAY in the VHH, a well-known hotspot forlater type could be referred to as non-templated nucleotidehypermutation (Milstein et al., 1998). These mutations ininsertions as often found at the VD or DJ joints. Thoughthe germline lead to the amino acid substitutions F27Ythe actual mechanism is still unknown, this observationand F29Y, two key-elements for the H1 loop conformationopens the possibility that the nucleotide addition process(Chothia et al., 1992). Further mutations of Y27 or Y29could occur during the rejoining of double strand DNAin the VHHs has two pronounced effects. Some mutationsbreaks of the VHHs.will lead to complete new loop conformations and therebyUnder any hypothesis, it is important to note that thewill expand the structural repertoire of the VHHs. OtherIns/Dels are not randomly distributed, but often occurmutations might lead to amino acid substitutions provokingnear or within the paratope. Therefore, they substantiallysubtle surface modifications that might improve the VHH–reshape the VHH loop structure. This is true for both theantigen fit. For these reasons, we propose to vary theseconventional VH and the VHH. However, in dromedary theamino acids in libraries of synthetic single domain VHHsincidence of off-size clones is much higher in the VHHor human VHs (Davies and Riechmann, 1996; Reiter et al.,cDNA than in the conventional VH cDNA (30 versus1999) to increase the potential repertoire and to search1.5%), indicating that the VHH is more prone to these

changes. Moreover, whereas in human most Ins/Dels were for more potent Ag binders.

928

Camel VH and VHH germline genes

Loo for help with the evolutionary analysis, and all colleagues for theirIn conclusion, the data presented here demonstrate thatbinders’ sequences. N.V.K. was supported by a VLIR-ABOS trainingthe Ag-binding repertoire of HCAb, derived from ~40grant. This work was granted by VLIR, VIB and FWO.germline VHHs, is largely diversified by specific mechan-

isms. These include the introduction of a variety ofinterloop disulfide bridges, the increased surface area of

Referenceshypervariable regions and a high rate of paratope reshap-Almagro,J.C., Hernandez,I., del Carmen,R. and Vargas-Madrazo,E.ing. The use of these specific mechanisms also results in

(1997) The differences between the structural repertoires of VH germ-novel paratopes different from those of the conventionalline gene segments of mice and humans: implication for the molecularAbs. This could explain the large proportion of HCAbsmechanism of the immune response. Mol. Immunol., 34, 1199–1214.

acting as competitive enzyme inhibitors. It is clear that Becker,R.S. and Knight,K.L. (1990) Somatic diversification ofparticular DNA signals with a high incidence in the immunoglobulin heavy chain VDJ genes: evidence for somatic gene

conversion in rabbits. Cell, 63, 987–997.VHHs trigger these somatic diversification processes. It isBerek,C., Berger,A. and Apel,M. (1991) Maturation of the immunepossible that the absence of the light chain provides the

response in germinal centers. Cell, 67, 1121–1129.freedom to make these changes possible and to allow theBrodeur,P.H. and Riblet,R. (1984) The immunoglobulin heavy chainVHH to explore new structures. variable region (Igh-V) locus in the mouse. I. One hundred Igh-V

genes comprise seven families of homologous genes. Eur. J. Immunol.,14, 922–930.Materials and methods

Cha,R.S. and Thilly,W.G. (1993) Specificity, efficiency and fidelity ofPCR. PCR Methods Appl., 3, S18–S29.Southern blots

Chothia,C., Boswell,D.R. and Lesk,A.M. (1988) The outline structureHigh MW genomic DNA was prepared from liver tissue of a singleof the T-cell αβ receptor. EMBO J., 7, 3745–3755.camel (Camelus dromedarius) from Morocco. Dromedary genomic DNA

(10 µg) was digested with BamHI, EcoRI or BamHI–EcoRI and Chothia,C. et al. (1989) Conformations of immunoglobulin hypervariableelectrophoresed in a 0.8% agarose gel. The DNA was transferred to regions. Nature, 342, 877–883.Hybond-N (Amersham) membranes. Two homologous VH probes of Chothia,C., Lesk,A.M., Gherardi,E., Tomlinson,I.M., Walter,G.,0.6 kb corresponding to the dromedary germlines VH and VHH (Nguyen Marks,J.D., Llewelyn,M.B. and Winter,G. (1992) Structural repertoireet al., 1998) were obtained by PCR and labelled with 32P (Radprime, of the human VH segments. J. Mol. Biol., 227, 799–817.Gibco-BRL). After hybridization, the membranes were washed at 65°C Davies,J. and Riechmann,L. (1995) Antibody VH domains as smallfor 15 min/wash (twice with 2� SSC, 0.1% SDS, once with 2� SSC, recognition units. Biotechnology, 13, 475–479.1% SDS, and finally with 0.1� SSC), and autoradiographed. Davies,J. and Riechmann,L. (1996) Single antibody domains as small

recognition units: design and in vitro antigen selection of camelized,Genomic DNA amplification, cloning and DNA sequencing human VH domains with improved protein stability. Protein Eng., 9,The primers used to amplify VH/VHH genes were derived from the VH 531–537.and VHH genomic sequences (Nguyen et al., 1998). One specific primer

Decanniere,K., Desmyter,A., Lauwereys,M., Ghahroudi,M.A.,of 26 nt corresponds to the upstream conserved Ig octamer sequenceMuyldermans,S. and Wyns,L. (1999) A single-domain antibody(VHOCTB: 5�-TCTATATATCTAGATGACATGCAAAT-3�). The otherfragment in complex with RNase A: non-canonical loop structuresprimer anneals at the FR3 sequence and starts from the Cys92 codonand nanomolar affinity using two CDR loops. Structure, 7, 361–370.(VHFR3F: 5�-ACAGTAATACATGGCCGTGTCCTC-3�). PCRs (50 µl)

Desmyter,A., Transue,T.R., Ghahroudi,M.A., Thi,M.H., Poortmans,F.,were performed on 0.5 µg liver genomic DNA with 2.5 U of Taq DNAHamers,R., Muyldermans,S. and Wyns,L. (1996) Crystal structure ofpolymerase (ROCHE) (30 cycles of 30 s at 94°C, 30 s at 55°C and 60 sa camel single-domain VH antibody fragment in complex withat 72°C with a final incubation of 10 min at 72°C). PCR products werelysozyme. Nature Struct. Biol., 3, 803–811.gel-purified, ligated into TA-PCRII cloning vector (Invitrogen), and

Devereux,J., Haeberli,P. and Smithies,O. (1984) A comprehensive set ofsubsequently transformed into DH5α cells. Randomly chosen clonessequence analysis programs for the VAX. Nucleic Acids Res., 12,with inserts of the expected size were sequenced from both ends. Two387–395.sets of PCR and cloning were carried out independently to exclude

Dufour,V., Malinge,S. and Nau,F. (1996) The sheep Ig variable regionpossible PCR errors.repertoire consists of a single VH family. J. Immunol., 156, 2163–2170.

Screening of a genomic library Ghahroudi,M.A., Desmyter,A., Wyns,L., Hamers,R. and Muyldermans,S.A camel liver genomic library (Nguyen et al., 1998) was screened by (1997) Selection and identification of single domain antibodyplaque hybridization with the same probes used in Southern blot fragments from camel heavy-chain antibodies. FEBS Lett., 414,experiment. The VH/VHH element from the putative positive phage clone 521–526.was amplified using the VHFR3F primer and a primer that corresponds Hamers-Casterman,C., Atarhouch,T., Muyldermans,S., Robinson,G.,to the leader-peptide sequence: VHLB 5�-GGCTGAGCTCGGTGGT- Hamers,C., Songa,E.B., Bendahman,N. and Hamers,R. (1993)CCTGGCT-3�. The phage-derived PCR fragments were cloned, and two Naturally occurring antibodies devoid of light chains. Nature, 363,clones were sequenced for each experiment. 446–448.

Hoogenboom,H.R. and Winter,G. (1992) By-passing immunisation.Camel V databases

Human antibodies from synthetic repertoires of germline VH geneGermline VH DNA sequences were aligned and grouped according tosegments rearranged in vitro. J. Mol. Biol., 227, 381–388.sequence identity. DNA sequences (600 bp) differing by no more than

Howell,M.D., Diveley,J.P., Lundeen,K.A., Esty,A., Winters,S.T.,four bases were considered identical as these differences might be dueCarlo,D.J. and Brostoff,S.W. (1991) Limited T-cell receptor β-chainto PCR or sequencing errors. The camel germline VH/VHH databaseheterogeneity among interleukin 2 receptor-positive synovial T cellscontains the remaining unique sequences. The camel cDNA data weresuggests a role for superantigen in rheumatoid arthritis. Proc. Natlcollected from the VH and VHH cDNA sequences available withinAcad. Sci. USA, 88, 10921–10925.our group.

Kabat,E.A., Wu,T.T., Perry,H.M., Gottesman,K.S. and Foeller,C. (1991)Sequences of proteins of immunological interest. US Public HealthSequence analysis

All DNA and protein sequence analyses were performed with version Services, NIH Bethesda, MD, Publication No. 91-3242.10 of the University of Wisconsin Genetics Computer Group package Klein,U., Goossens,T., Fischer,M., Kanzler,H., Braeuninger,A.,(Devereux et al., 1984). Rajewsky,K. and Kuppers,R. (1998) Somatic hypermutation in normal

and transformed human B cells. Immunol. Rev., 162, 261–280.Kleinfield,R., Hardy,R.R., Tarlinton,D., Dangl,J., Herzenberg,L.A. andAcknowledgements

Weigert,M. (1986) Recombination between an expressedimmunoglobulin heavy-chain gene and a germline variable geneWe thank C.Bouton, M.De Kerpel, Y.Hou and L.M.Tam for technical

assistance, Dr K.Decanniere and D.Maes for discussions, Dr W.van der segment in a Ly 1� B-cell lymphoma. Nature, 322, 843–846.

929

V.K.Nguyen et al.

Knight,K.L. (1992) Restricted VH gene usage and generation of antibody Vu,K.B., Ghahroudi,M.A., Wyns,L. and Muyldermans,S. (1997)Comparison of llama VH sequences from conventional and heavydiversity in rabbit. Ann. Rev. Immunol., 10, 593–616.chain antibodies. Mol. Immunol., 34, 1121–1131.Komori,T., Minami,Y., Sakato,N. and Sugiyama,H. (1993) Biased usage

Wagner,S.D. and Neuberger,M.S. (1996) Somatic hypermutation ofof two restricted VH gene segments in VH replacement. Eur. J.immunoglobulin genes. Ann. Rev. Immunol., 14, 441–457.Immunol., 23, 517–522.

Weinstein,P.D., Anderson,A.O. and Mage,R.G. (1994) Rabbit IgHLauwereys,M., Ghahroudi,M.A., Desmyter,A., Kinne,J., Holzer,W., Desequences in appendix germinal centers: VH diversification by geneGenst,E., Wyns,L. and Muyldermans,S. (1998) Potent enzymeconversion-like and hypermutation mechanisms. Immunity, 1, 647–659.inhibitors derived from dromedary heavy-chain antibodies. EMBO J.,

Wilson,P.C., de Bouteiller,O., Liu,Y.J., Potter,K., Banchereau,J.,17, 3512–3520.Capra,J.D. and Pascual,V. (1998) Somatic hypermutation introducesMartin,A.C. and Thornton,J.M. (1996) Structural families in loops ofinsertions and deletions into immunoglobulin V genes. J. Exp. Med.,homologous proteins: automatic classification, modelling and187, 59–70.application to antibodies. J. Mol. Biol., 263, 800–815.

Winter,G., Griffiths,A.D., Hawkins,R.E. and Hoogenboom,H.R. (1994)Melek,M., Gellert,M. and van Gent,D.C. (1998) Rejoining of DNA byMaking antibodies by phage display technology. Ann. Rev. Immunol.,the RAG1 and RAG2 proteins. Science, 280, 301–303.12, 433–455.Milstein,C., Neuberger,M.S. and Staden,R. (1998) Both DNA strands of

Woolven,B.P., Frenken,L.G., van der Logt,P. and Nicholls,P.J. (1999)antibody genes are hypermutation targets. Proc. Natl Acad. Sci. USA,The structure of the llama heavy chain constant genes reveals a95, 8791–8794.mechanism for heavy-chain antibody formation. Immunogenetics, 50,Muyldermans,S., Atarhouch,T., Saldanha,J., Barbosa,J.A. and Hamers,R.98–101.(1994) Sequence and structure of VH domain from naturally occurring

Yelamos,J., Klix,N., Goyenechea,B., Lozano,F., Chui,Y.L.,camel heavy chain immunoglobulins lacking light chains. ProteinGonzalez,F.A., Pannell,R., Neuberger,M.S. and Milstein,C. (1995)Eng., 7, 1129–1135.Targeting of non-Ig sequences in place of the V segment by somaticNeuberger,M.S., Ehrenstein,M.R., Klix,N., Jolly,C.J., Yelamos,J.,hypermutation. Nature, 376, 225–229.Rada,C. and Milstein,C. (1998) Monitoring and interpreting the

intrinsic features of somatic hypermutation. Immunol. Rev., 162,Received November 16, 1999; revised January 4, 2000;107–116.

accepted January 5, 2000Nguyen,V.K., Muyldermans,S. and Hamers,R. (1998) The specificvariable domain of camel heavy-chain antibodies is encoded in thegermline. J. Mol. Biol., 275, 413–418.

Nguyen,V.K., Hamers,R., Wyns,L. and Muyldermans,S. (1999) Loss ofsplice consensus signal is responsible for the removal of the entireCH1 domain of the functional camel IgG2a heavy-chain antibodies.Mol. Immunol., 36, 515–524.

Novotny,J. (1991) Protein antigenicity: a thermodynamic approach. Mol.Immunol., 28, 201–207.

Parng,C.L., Hansal,S., Goldsby,R.A. and Osborne,B.A. (1996) Geneconversion contributes to Ig light chain diversity in cattle. J. Immunol.,157, 5478–5486.

Ramsden,D.A., McBlane,J.F., van Gent,D.C. and Gellert,M. (1996)Distinct DNA sequence and structure requirements for the two stepsof V(D)J recombination signal cleavage. EMBO J., 15, 3197–3206.

Reiter,Y., Schuck,P., Boyd,L.F. and Plaksin,D. (1999) An antibody single-domain phage display library of a native heavy chain variable region:isolation of functional single-domain VH molecules with a uniqueinterface. J. Mol. Biol., 290, 685–698.

Reth,M., Gehrmann,P., Petrac,E. and Wiese,P. (1986) A novel VH toVHDJH joining mechanism in heavy-chain-negative (null) pre-B cellsresults in heavy-chain production. Nature, 322, 840–842.

Reynaud,C.A., Anquez,V., Grimal,H. and Weill,J.C. (1987) Ahyperconversion mechanism generates the chicken light chainpreimmune repertoire. Cell, 48, 379–388.

Reynaud,C.A., Dahan,A., Anquez,V. and Weill,J.C. (1989) Somatichyperconversion diversifies the single Vh gene of the chicken with ahigh incidence in the D region. Cell, 59, 171–183.

Schiff,C., Corbet,S. and Fougereau,M. (1988) The Ig germline generepertoire: economy or wastage? Immunol. Today, 9, 10–14.

Schroeder,H.W.J., Hillson,J.L. and Perlmutter,R.M. (1990) Structure andevolution of mammalian VH families. Int. Immunol., 2, 41–50.

Sinclair,M.C., Gilchrist,J. and Aitken,R. (1997) Bovine IgG repertoireis dominated by a single diversified VH gene family. J. Immunol.,159, 3883–3889.

Spinelli,S., Frenken,L., Bourgeois,D., de Ron,L., Bos,W., Verrips,T.,Anguille,C., Cambillau,C. and Tegoni,M. (1996) The crystal structureof a llama heavy chain variable domain. Nature Struct. Biol., 3,752–757.

Stewart,A.K., Huang,C., Long,A.A., Stollar,B.D. and Schwartz,R.S.(1992) VH-gene representation in autoantibodies reflects the normalhuman B-cell repertoire. Immunol. Rev., 128, 101–122.

Sun,J., Kacskovics,I., Brown,W.R. and Butler,J.E. (1994) Expressedswine VH genes belong to a small VH gene family homologous tohuman VHIII. J. Immunol., 153, 5618–5627.

Tomlinson,I.M., Walter,G., Marks,J.D., Llewelyn,M.B. and Winter,G.(1992) The repertoire of human germline VH sequences reveals aboutfifty groups of VH segments with different hypervariable loops. J. Mol.Biol., 227, 776–798.

Tonegawa,S. (1983) Somatic generation of antibody diversity. Nature,302, 575–581.

930